2. Suspension preparation and chemical analysis
2.2 Chemical analysis
2.2.1 Recovery in different matrices
Based on the results of the analytic and discussion with the co-project FKZ 3709 65 416 and 3709 65 418 (both coordinated by IME) following digestion methods were used (Table 6):
Table 6: Overview of the applied digestion methods for the different matrices.
Stock suspensions (DI water and SDW)
Laboratory sewage treatment plant
- effluent
Laboratory sewage treatment plant
- sludge
Soil - eluate
Soil – matrix
HCl, HNO3, HF √ √ √ √ -
H2SO4
microwave - - - √
Further to the validation of the principal analytical method the identification and calculation of the Titanium in the different matrices are crucial information for further data interpretation and conclusion. Therefore an evaluation of the different digestion methods for different matrices was conducted to test for the recovery of TiO2 nanomaterials.
Stock suspensions in synthetic drinking water – Laboratory sewage treatment plant The experiments in the laboratory sewage treatment plant (LSTP) were conducted with different concentrations of P25 – 100 mg/L, 50 mg/L and 10 mg/L (for details see chapter 3.1 Laboratory sewage treatment plant – OECD 303A). For the experiment P25 suspensions were prepared in synthetic drinking water (SDW), which was added to the system every 24 h.
Before the sewage treatment plant tests were performed the recovery of three different P25 concentrations in synthetic drinking water was analysed (Figure 19).
The recovery analysis revealed a concentration dependent linear correlation (R² = 0.99) with lower recovery rates for stock suspensions of higher concentration. Overall an average recovery of 92 ± 3% (arithmetic mean) and 88% for the slope of the stock suspension in synthetic drinking water was determined (Figure 19). Therefore an average recovery of 90%
for Ti detection, based on arithmetic mean and slope, from particulate TiO2 is derived.
0%
20%
40%
60%
80%
100%
120%
10 mg/L 50 mg/L 100 mg/L
recovery in %
TiO2 stock suspension
y = 0.884x + 1.0703 R2 = 0.9997
0 20 40 60 80 100 120
0 20 40 60 80 100 120
TiO2stock suspension in mg/L measured TiO2 in mg/L
Figure 19: Recovery (upper and lower figure) of the TiO2 content from the stock suspensions in synthetic drinking water using HCl, HNO3, HF digestion. Error bars = min and max; n = 2.
A concentration dependent linear correlation of the recovery rate was found with an average recovery rate of 90% for SDW. The variability between the results of the different concentrations was within the range of the uncertainty of the detection method.
Stock suspension in deionised water – soil experiments
All three different TiO2 materials (solid) had to be suspended in deionised water (see also chapter 3.2 Leaching experiments in soil columns – OECD 312) before they could be used in the leaching experiments in soils as well as for the adsorption experiments. Therefore the recovery rate for Titanium in deionised water was tested for three different concentrations for each TiO2 material. A linear correlation between stock suspension and measured concentration with a R² = 0.99 was detected for all materials tested (Figure 20) with slopes ranging from 1.03 to 1.23.
The lowest recovery rate (64% - 80%) was always determined for the lowest Titanium concentration of 10 mg/L while the highest recovery rates were found for the concentrations of 1000 mg/L and 5000 mg/L (103% - 149%). This indicates a concentration dependent and linear recovery rate for all materials (Figure 21).
A material dependent recovery rate is also indicated by the data, which vary between arithmetic average recovery rates of 92% and 114%.
The coating of the UV Titan M262 or the different crystal forms of the materials could be the reasons for this finding: PC105 100% anatase; P25 86% anatase, 14% rutile; UV Titan M262 100% rutile. The lowest recovery was detected for the rutile form and the maximum recovery for the anatase crystal form.
y = 1.2327x + 73.549 R2 = 0.998
y = 1.1702x - 30.416 R2 = 0.9998
0 1000 2000 3000 4000 5000 6000 7000
0 1000 2000 3000 4000 5000 6000
TiO2 stocksupension in mg/L measured TiO2 in mg/L
PC105 P25
UV Titan M262
y = 1.0381x + 25.773 R2 = 0.9994
Figure 20: Linear correlation of TiO2 concentrations in DI water as prepared and measured using HCl, HNO3, HF digestion. Stock suspension concentrations 10 mg/L, 100 mg/L, 1000 mg/L and 5000 mg/L. Error bars covered by the symbols; n = 3.
0%
20%
40%
60%
80%
100%
120%
140%
160%
10 mg/L 100 mg/L 1000 mg/L 5000 mg/L TiO2stock suspension
recovery in %
PC105
0%
20%
40%
60%
80%
100%
120%
140%
160%
10 mg/L 100 mg/L 1000 mg/L 5000 mg/L TiO2 stock suspension
recovery in %
P25
0%
20%
40%
60%
80%
100%
120%
140%
160%
10 mg/L 100 mg/L 1000 mg/L 5000 mg/L TiO2stock suspension
recovery in %
UV Titan M262
Figure 21: Ti recovery rate for the three different TiO2 nanomaterials in DI water using HCl, HNO3, HF digestion. Four concentrations: 10 mg/L, 100 mg/L, 1000 mg/L and 5000 mg/L. PC105 (upper figure), UV Titan M262 (central figure) and P25 (lower figure). Filled part of the bar = measured recovery value.
In summary a material, concentration dependent and linear correlated (R² = 0.99) recovery was determined. The overall arithmetic average recovery was determined to be 105% ±
42%. The slopes for all materials ranged between 103% and 123 % with negligible Y-axis values. With the concentrations measured mainly at the lower end of the concentration range tested here an average recovery of 100% in DI water is assumed.
water is assumed. The high variance with values differencing by 0% have to be noted.
Taking into account concentrations measured in project samples, a recovery rate (average) of 100% in DI
5
Recovery in sewage sludge
In the laboratory sewage treatment plant experiment, the sewage sludge beside the effluent was analysed for its TiO2 content to conduct the mass balance for the P25 in this system. To derive information about the recovery rates, sewage sludge of the reference system was spiked with different TiO2 stock suspensions. A linear recovery with an R² = 0.98 of the added P25 was identified with an average arithmetic recovery rate of 58% and 37% based on the slope was determined (Figure 22).
0%
20%
40%
60%
80%
100%
120%
8.7 mg/L 43.5 mg/L 89 mg/L
TiO2in sludge
recovery in %
y = 0.3651x + 3.4841 R2 = 0.9809
0 10 20 30 40 50 60 70 80 90 100
0 10 20 30 40 50 60 70 80 90 100
TiO2 sewage sludge in mg/L measured TiO2 sewage sludge in mg/L
Recovery (upper and lower figure) of the TiO2 content from spiked sewage sludge aliquots using HCl, HNO3, HF digestion. F
Figure 22:
illed part of the bar = measured recovery value, error bars = min and max; n = 2.
two samples spiked with the highest An unsatisfying recovery rate of 39% and 42% for the
concentration (43.5 mg/L and 89 mg/L) was detected.
The main part of the sewage sludge (about 75%) consists of organic carbon. The inorganic carbon is further reduced during the test through the removal of surplus sludge. Therefore an
influence of the matrix on the recovery rate or interference with the analysis devices seems implausible (complete chemical digestion of the organic carbon).
The sample preparation could have an effect on the recovery rate, because a homogeneous spiking of the sewage sludge can be a problem, due to agglomeration of the particles in suspension by contact with the sludge, which could lead to an inhomogeneous distribution in
of the stock suspensions in synthetic drinking water for P25 a recovery rate of 90% may be a u
the covery rate of the stock suspension in synthetic drinking water of 90% may be taken
interpretation of the mass balance of the sewage plant experiments.
the sample. The unsatisfying recovery rate of 39% and 42% of the samples spiked with the highest concentration (43.5 mg/L and 89 mg/L) can be an indication for this.
No clear explanation on the variance in the recovery rate can be given at this time. Based on the implausibility of this finding, the plausibility of the mass balance for the sewage plant experiments and the results for the stock suspensions we assume that the spiked sewage sludge could not be used and were treated as outliers. Based on the recovery rate
ss med for the interpretation of the mass balance in the sewage plants experiments.
An insufficient and implausible recovery was detected for samples with the two highest concentrations. Based on the results of the recovery rate for the other matrices and in the context of the chemical analysis, these results were treated as outliers and re
into account for
Recovery in soils
Two reference soils (reference soil materials SRM 2709a - San Joaquin Soil, NIST and the BCR 142 - light sandy soil, JRC) were analysed for their Ti content for a first method validation and further data comparison ( ), as the analytical procedure was described in chapter 2.2
Figure 23
. A comparison measurement with the co project FKZ 3709 65 416 and 3709 65 418 (lead by Fraunhofer IME) of different soil samples was conducted (results in Annex II).
0%
BCR - 142 San Joaquin - 2709a
Recovery of Ti content of two reference soils - using H2SO4 micro
20%
40%
60%
80%
100%
120%
recovery in %
Figure 23: wave digestion.
100% is the value for the reference Ti content of the soils. Filled part of the bar = measured recovery value, error bars = min and max; n = 6.
levance for the conducted experiments since only the recovery rate of added l.
It is expected that with the used H2SO4 microwave digestion, not all silicates and minerals of the soil were digested, therefore explaining the lower recovery rate of on average 72 ± 4%.
This is not of re
TiO2, which was tested with spiked soils, is of relevance for the mobility analysis of the nanomateria
Uncorrected values were used for the following analysis, unless specifically indicated differently.
The recovery of the three different TiO2 materials in the soil matrix was tested in different soils. First two soils (A01 and G03) were spiked with P25 to derive information about the recovery rate of the added P25 (Figure 24).
0%
spiked G03 spiked A01
Recovery of P25 (40,000 mg/kg) in soil type G0
20%
40%
60%
80%
100%
120%
recovery in %
Figure 24: 3 - Eutric Cambisol and A01 -
Dystric Cambisol (H2SO4 microwave digestion. Filled part of the bar = measured recovery value, error bars = min and max; n = 2).
tration dependent effect has an influence on the recovery rate. Analogue PC105 and UV Titan M262 were tested (Figure 25) by using lower concentrations (1 mg/kg and 5 mg/kg).
The recovery rate of P25 (40,000 mg/kg) was measured, to make sure that no matrix (different mineral content) or concen
0%
20%
40%
60%
80%
100%
120%
140%
160%
180%
200%
1 mg/kg 5 mg/kg 1 mg/kg 5 mg/kg
recovery in %
spiked A04 spiked A01
0%
20%
40%
60%
80%
100%
120%
140%
160%
180%
200%
1 mg/kg 5 mg/kg 1 mg/kg 5 mg/kg
recovery in %
spiked A04 spiked A01
PC105
UV Titan M262
Figure 25: Recovery of PC105 (upper figure) and UV Titan M262 (lower figure) using a H2SO4 microwave digestion for soil type A04 – Gleyic Podsol and soil type A01 - Dystric Cambisol. Filled part of the bar = measured recovery value, error bars = min and max; n = 2.
The recovery rate of the two tested soils was 72 ± 4%. The low recovery can be due to the chemical digestion method chosen.
The recovery of P25 was 82 ± 12% for both soils with a higher recovery of 92% for soil A01.
The average recovery for PC105 was 132% ± 40% and for UV Titan M262 96% ± 27% with a higher recovery in soil A01 with 113% and a lower recovery in soil A04 with 80%.
The recovery tests in the stock suspension showed a material dependent recovery with a higher recovery for PC105 and a lower recovery for UV Titan M262. Similar results were found for the spiked soils. If this was considered it could be that the type of the material could affect the recovery in soils, which we took into account for further analysis.
The recovery of the three different TiO2 nanomaterials in a complex soil matrices showed a higher variance compared with the recovery of the stock suspension. The higher variance could be caused by matrix effects and the difficulty of the preparation of homogeneous samples of the spiked soils.
A rough trend was shown with the highest recovery rate for PC105 (132% ± 40%), than UV Titan M262 (96% ± 27%) and than P25 (82% ± 12%).
The recovery rate of the different TiO2 nanomaterials in the complex soil matrix is acceptable against the background of the natural high amount of Ti in natural soils and was within the range of the uncertainty of the detection method (chapter 2.2).
Overall a recovery rate of 100 % for TiO2 in soil matrices is deduced from all values.
Conclusion
The highest variance of the analysis was detected for soils, the most complex matrix tested here. This finding may be due to the matrix itself or by the preparation of the spiked sludge or soil (inhomogeneous dispersion of the nanomaterials in the samples – local accumulation). A nanomaterial dependent recovery is indicated (Table 7):
- PC105 highest recovery for the liquid and the soil matrix,
- UV Titan M262 medium recovery the liquid matrix and lowest recovery in soil matrix, - P25 the lowest recovery in the liquid and a medium recovery in soil matrix.
Table 7: Recovery rates in percent of the tested TiO2 nanomaterials in different matrices.
SDW – suspension
average (min; max)
DI – suspension
average (min; max)
Sludge average (min; max)
spiked soil A01 average (min; max)
spiked soil A04 average (min; max)
spiked soil G03 average (min; max)
P25 92
(89; 95)
101 (80; 117)
58 (39; 92)
92
(92; 93) - 71
(63; 79)
PC105 - 121
(64; 149) - 132
(92; 170) 133
(108; 159) - UV
Titan M262
- 94
(64; 115) - 113
(96; 130)
80
(78; 82) -
The overall recovery for the (liquid) samples for the LSTP experiment was 90%.
The overall recovery for the stock suspensions for the soil experiments was 100%, with a trend of material dependent recovery: PC105 (121% ± 35%), P25 (101% ± 18%) and UV Titan M262 (94% ± 27%).
The overall recovery of the three added TiO2 materials in soils was 100%, with a trend of material dependent recovery: PC105 (132% ± 40%), UV Titan M262 (96% ± 27%) and P25 (82% ± 12%).
The material dependency was used for the interpretation of the mass balance in the soil column experiments, but the values were not corrected for this recovery rates.